On my list of potential topics to blog about in the last few days was Megan McArdle's excellent post in which she advises people to save more. I've gone after her here, here, and here and so I like to balance things because she really is a very sharp, informed, analytic blogger.
Yes, Arnold is right that if saving doesn't lead to more capital formation, then one person's saving is another person's dissaving. (Arnold's claim that it's zero sum doesn't follow, by the way, because, as commenter Fralupo pointed out, people gain from trade and in this case they are trading on different discount rates.)
Commenter Bob Murphy implicitly points out the problem: the assumption that an increase in saving nationwide would not lead to an increase in capital formation is highly implausible. This is, essentially, a Klingian paradox of thrift.